STRUCTURED MODELS AND DYNAMIC SYSTEMS ANALYSIS: THE INTEGRATION OF THE IDEF0/IDEF3 MODELING METHODS AND DISCRETE EVENT SIMULATION Larry Whitman Adrien Presley Brian Huff Automation & Robotics Research Institute Division of Business and Accountancy The University of Texas at Arlington Truman State University 7300 Jack Newell Boulevard South 100 East Normal Fort Worth, Texas 76118, U.S.A. Kirksville, Missouri 63501, U.S.A. · To analyze and design the enterprise and its ABSTRACT processes prior to implementation · To help reduce complexity The role of modeling and simulation is receiving much · To communicate a common understanding of the press of late. However, the lack of practice in employing system a link between the two is alarming. A static model is · To gain stakeholder buy-in used to understand an enterprise or a system, and · To act as a documentation tool for ISO 9000, TQM, simulation is used for dynamic analysis. Generally, most Concurrent Engineering, and other efforts. models are considered static, whereas simulation is really A primary thrust of this research is to determine the a dynamic model. Static models are useful in achieving feasibility of using a single master static model of the understanding of the enterprise. Simulations are useful in enterprise for multiple purposes. Previous research has analyzing the behavior of the enterprise. presented a single suite perspective (Lingineni, Caraway Most enterprises develop and even maintain multiple et al. 1996) and uses custom developed software (Harrell types of models for different purposes. If a single model and Field 1996). This research uses commercial products can be used to drive other modeling purposes, then exclusively. For the static model, the research uses two model maintenance and development could be reduced. specific methods for static representation, IDEF0 and This paper describes the procedure necessary to use a IDEF3. Two concerns were: (1) the amount of required static representation as the primary input for an animated change of the IDEF methodology, and (2) the amount of simulation. It presents the additional steps necessary to additional annotation to the IDEF model required by the annotate a static model for input to a dynamic model. IDEF tool. Further explanation of these concerns Two commercial suites, WorkFlow ModelerÔ to follows. ServiceModelÔ and ProSimÔ to WITNESSÔ, are IDEF is a rigorous methodology. The reason for the compared and contrasted based on the respective ease of rigor is to ensure a robust and complete representation. conversion from the static model to the dynamic model. As part of this rigor, a thorough review process is used. Any user who purchases these products can follow the The review cycle is enhanced by the rigid IDEF syntax. steps described in this paper for either of these product The syntax for IDEF is very explicit. A concern in suites to generate a simulation from a static model. utilizing an IDEF model to create a simulation is that the Finally, some general observations of using an existing IDEF method would have to be compromised to enable IDEF (0 or 3) model to create a working simulation are the creation of a simulation directly from the IDEF presented along with conclusions. model. However, it should also be noted that there are certain characteristics of IDEF modeling that have 1 INTRODUCTION become considered standard practice, yet they are not a strict IDEF syntactic rule. This research found that the The commonly accepted definition is that a model is a only changes to the methodology required were in these representation of reality. Generally, details that are time-honored traditions, whereas no actual IDEF rule of unnecessary are not included. The typical uses (Nathan syntax was broken. and Wood 1991) (Snodgrass 1993) (Reimann and Sarkis 1996) of modeling are: Integration of the IDEF0/IDEF3 Modeling Methods and Discrete Event Simulation 519 Integrated Modeling and Simulation Environment IDEF Tool Simulation Tool IDEF Methodology Simulation Specific Information Figure 1: Integrated Modeling and Simulation Environment Another concern was the amount of additional possible flow paths of objects through a system. This annotation required by the tool to drive the simulation. information is very helpful in determining what items Figure 1 shows how the IDEF methodology, the IDEF participate in the process and the functions performed by tool, and the simulation data interact. Note the the system. Although static representations can indicate distinction between the IDEF methodology and the IDEF the allowable system behaviors, they cannot depict the tool. The tool is an implementation of the methodology range of time-variant behavior generated as a result of itself. Tools sometimes have limitations which the resource availability or the number of items flowing methodology does not require. The opposite is also true through the process. To adequately predict the in that the methodology sometimes restricts a certain performance characteristics of dynamic systems, the characteristic, yet the tool does not restrict this feature. time-variant behaviors of the system must be able to be Therefore, it is important to note the differences between defined and represented. the limitations of methodologies and tools. A primary research task was to validate the need for additional 2.2 Dynamic population of data inside the simulation itself. The goal is to eliminate the need for the user to add any Dynamic representations of systems attempt to capture information inside the simulation tool. This will facilitate and describe the behavior of the system over time under the master model concept discussed later. different operating conditions. For the purposes of this The paper will first discuss the types of models and paper, we are referring to discrete-event simulation as explain the differences between static and dynamic the dynamic system model. Although the static system models. Next, we will provide an overview of the IDEF representations are capable of providing the vast methodology that includes an explanation of the tools majority of the information needed to construct a used. Then, the static model creation process is dynamic systems model, they do not possess the mechanisms needed to enact the process behavior discussed, including a description of the sample process constraints defined in their representations. Discrete- used and the process of creating the static models. A event simulation tools, in contrast, are capable of description is then given of the method used to convert executing sets of system behavior roles and tracking the the static model to a dynamic model. Some general system’s transition through a series of states. In this observations are made and future directions are then manner, a dynamic model can provide information about presented. the state of the system at a given instance in time or can generate performance measures of the system over a 2 TYPES OF MODELS given period of time. Dynamic models can be used iteratively to study system behavior under different In this section, we discuss the two types of models: static operating conditions. Subtle changes in resource and dynamic, and explain the five different views of a availability or system loading can have dramatic effects model. on the performance of the system. This range of potential 2.1 Static behaviors is very difficult to represent with a static system model. Static models attempt to provide a static representation of dynamic systems. Static models generally portray the 520 Whitman, Huff, and Presley 2.3 Views 3.1 IDEF0 Previous work in the development of architectures by the There are five elements in the IDEF0 functional model Automation & Robotics Research Institute (Presley, as shown in Figure 2. The activity (or function) is Huff, Liles 1993) describes a five-view approach. The represented by the boxes; inputs are represented by the Business Rule (or Information) View defines the entities arrows flowing into the left hand side of an activity box; managed by the enterprise and the rules governing their outputs are represented by arrows flowing out the right relationships and interactions. The Activity View defines hand side of an activity box; the arrows flowing into the the functions performed by the enterprise (what is done) top portion of the box represent constraints or controls while the Business Process View defines a time- on the activities; and the final element represented by sequenced set of processes (how it is done). The arrows flowing into the bottom of the activity box are the resources and capabilities managed by the enterprise are mechanisms that carry out the activity (Marca and defined in a Resource View. Finally, the Organization McGowan 1988, Mayer 1992). View is used to define how the enterprise organizes itself Constraint and the set of constraints and rules governing how it manages itself and its processes. This does not, however, mean that all these views Input Perform Output must be present in all models. A model is an abstract Activity representation of reality which should exclude details of the world which are not of interest to the modeler or the ultimate users of the model. Models are developed to Mechanism (Resource) answer specific questions about the enterprise. This research focuses specifically on the need for analysis of Figure 2: IDEF0 Nomenclature resource constraints and process flows. 3.2 IDEF3 3 OVERVIEW OF IDEF MODELS The IDEF3 Process Description Capture Method IDEF (Integration DEFinition) was developed by the (Mayer, Painter, deWitte 1992) consists of process flow U.S. Air Force’s Integrated Computer Aided diagrams and elaboration diagrams. Only the process Manufacturing (ICAM) project in the late 1980’s. There flow diagrams were used to feed the simulation model. are many different IDEF methods. Each method is useful IDEF3 uses a rigid syntax that eliminates model for describing a particular perspective of an enterprise. ambiguity. The basic elements of IDEF3 process The major IDEF methods in use are functional or activity descriptions used in this research are Unit of Behaviors modeling (IDEF0), information modeling (IDEF1), data (UOBs), Junctions, and Links.